Is your llms.txt AI-ready?
Validate your llms.txt against the emerging specification: structure, content, links, entity definitions, and use policy.
Optional for the engines, useful for the file itself.
llms.txt is a proposed file that summarizes your website for language models: a title, a short description, and sections of links to your key pages. It is optional: Google says Google Search ignores it, and no major AI engine documents using it to choose citations (checked September 2026).
This validator checks your file against Radar's rubric, which also scores entity definitions and a use policy, two sections that are not part of the llms.txt proposal.
What we validate.
6 checks, each with a plain-English reason it matters. Every one maps to a line in the report.
Structure & Formatting
H1 title, blockquote summary, markdown headings, and clean formatting.
Content Sections
Company info, products/services, descriptive content beyond just links.
Links & URLs
URL count, distribution across sections, internal vs external links.
Entity Definitions
“Is a/provides/offers” language that helps AI build entity associations.
Use Policy
Citation guidance, attribution rules, and contact information.
Completeness
Word count depth (no credit past 150,000 characters, about the largest 5% of llms.txt files we sampled), llms-full.txt bonus, and section variety.
How We Score (And Why It Matters)
The Crawl Checker does a shallow llms.txt check (4 binary signals, 20 pts). This Validator goes deep: section-by-section analysis, link extraction, entity detection, word count analysis, spec compliance, and content quality.
Built on our own implementation
We built a dynamic, entity-aware llms.txt for pixelmojo.io that auto-maps from our knowledge graph. This validator reflects what we learned about what makes an llms.txt effective for AI discovery.
Want the full picture across four AI engines?
Radar orchestrates 13 AI visibility audits in staged batches: crawl access, llms.txt, structured data, brand citations, and more. It then surfaces cross-tool conflicts and ships a prioritized action plan for your domain.
Run it on your site. See what AI can read.
Free, no account, result on this page. When you want the whole picture, Radar runs all 13 audits in one pass.